2 Basic Workflows - Vector

Let’s assume that we want to make some kind of a map, be it because we actually need a map (i.e. for a great book on European foreign policy), or maybe just to give some spatial data some kind of a cognitive canvas. As geodata tends to take up non-trivial amounts of resources (bandwidth, memory, CPU), it could be useful to focus on only a particular geographic extent. This is where I learned to love the concept of a bounding box (BBOX). Instead of literally downloading the whole world (thereby straining someone else’s bandwidth / server capacity) and then running costly query/filter operations, we can easily define bounds first.

If you’re spatially literate and therefore easily can spot whether you need Lat/Lon or Lon/Lat, or what projection / CRS your favourite BBOX-providing workflow offers, you probably can skip this and just define your BBOX manually with something like bbox <- c(xmin, ymin, xmax, ymax). I tend to struggle with this (even when using web tools such as http://bboxfinder.com), because I either mix up Lat/Lon or underestimate the spatial extent of my geodata or whatever. This is why I think that these three approaches below might be helpful to others, too.

2.1 Basemap Bounding Box: Three Approaches

So first of all we want to decide on the extent of the basemap. Since this is a recurrent task but with varying parameters, I’m offering three different approaches, based on

1. the bounding box (BBOX) of a spatial object (i.e. Admin 1 level Federal state|s)
1. the bounding box of spatial data (i.e. the geocoded cycling data)
1. a hand-drawn rectangle turned into a BBOX

2.1.1 Bounding Box: Object-based

Here, we start with our spatial object of reference. I prefer to work with Open Access / Public Domain data (and not Google Maps / Bing et al.), so let’s fetch some vector data from the Natural Earth project - but as class = Simple Feature instead of Esri’s proprietary (and less tidyverse-friendly) Shapefile format.

For your use case, consult the vignette for rnaturaleath::ne_download() and the Natural Earth website. For the sake of resource efficiency I also recommend to download the data once and then to store it locally with save/saveRDS for future use. As I work a lot with Natural Earth data, I have mounted a network folder D:/GIS/ with all the raster and vector data I’ve downloaded so far.However, I do not recommend download the “Download all X themes” file, since this gives you a) shapefiles instead of Simple Features, and b) UTF-8 encoding issues* for the feauture labels, depending on when the particular theme was parsed the last time.

substates10 <- rnaturalearth::ne_download(scale = 10, type = "admin_1_states_provinces", category = "cultural", returnclass = "sf")

# saveRDS(substates10, file = str_c(here::here("data", "GIS_workflows", "/"), "substates10.rds"))
# substates10 <- readRDS(file = str_c(here::here("data", "GIS_workflows"), "/", "substates10.rds"))

berlin_sf <- substates10 %>% filter(name == "Berlin")

Next, we’ll use the amazing sf package to fetch Berlin’s bounding box.

berlin_bbox <- sf::st_bbox(berlin_sf)

So… wanna have a quick glimpse at Berlin’s bounding box? Here’s all that it takes with the game-changing mapview package by Tim Appelhans et al.:

berlin_bbox %>% mapview::mapview()

Caveat: To keep memory and CPU load low for everyone browsing this post, I mostly will display the output of interactive widgets as .jpg.

If you want to set the “CartoDB Dark Matter” basemap as your default (or any other of the themes supported by Leaflet ), you might want to add this setting to your .Rprofile:

mapviewOptions(basemaps = c("CartoDB.DarkMatter", "CartoDB.Positron", "Esri.WorldImagery", "OpenStreetMap", "OpenTopoMap"))

And since mapview() is cool, we clan plot and inspect our Berlin object and the BBOX at the same time. For this we’re going to convert the BBOX object into a regular simple feature object with sf::st_as_sfc on the fly:

mapview::mapview(list(sf::st_as_sfc(berlin_bbox), berlin_sf))

This box would probably be too narrow for further static editing (i.e. for a print map), so it would be cool if we could simply increase the BBOX’s extent. Since the coordinates in the bbox are stored as a numeric vector c(xmin, ymin, xmax, ymax), we can easily expand the bbox by providing another vector of length = 4 and see if the extent is better on the fly:

(berlin_bbox + c(-0.1, -0.1, 0.1, 0.1)) %>%
  # sf::st_as_sfc() %>% 
  mapview::mapview()

That seems generous enough. So let’s preserve this box for later.

berlin_bbox <- berlin_bbox + c(-0.1, -0.1, 0.1, 0.1)

2.1.2 BBOX from data

Another approach to get a reasonable bounding box is to calculate the BBOX based on your geodata.

As mentioned above, I’m going to use the freshly released cycling data from the Berlin-based Tagespiegel DDJ / Innovation Lab for the final use case. The data has been collected as part of the sensor-based #radmesser project, where Lab leader Hendrik Lehmann’s team of journalists and developers equipped 100 (!) cyclists with close-range sensors to measure the distance of passing-by cars and trucks on Berlin’s roads. Goal: Demonstrate the at-risk status of cyclists in Germany’s capital.

Make sure to check out the award-winning project’s website and the project’s repo.

It’s probably worth the remark that getting the data into R is as simple as sf::st_read(URL)…

# License: ODC-By v1.0/Tagesspiegel Radmesser/https://radmesser.de
# cf. https://github.com/tagesspiegel/radmesser/blob/master/opendata/LICENSE.md
berlin_bike <- sf::st_read("https://github.com/tagesspiegel/radmesser/blob/master/opendata/detailnetz_ueberholvorgaenge.geo.json?raw=true")

# saveRDS(berlin_bike, file = str_c(here::here("data", "GIS_workflows", "/"), "berlin_bike.rds"))
# berlin_bike <- readRDS(file = str_c(here::here("data", "GIS_workflows"), "/", "berlin_bike.rds"))

However, st_read silently drops the “stats” field (which is legit, since a GeoJSON feature is defined as geometry + properties only) which contains the single measurements. Or rather: I have no 1-liner idea how to preserve it, be it in QGIS, with jsonlite, or by transforming with http://geojson.io. Fortunatly, the Radmesser-team also provides a CSV with the measurements and a column with the respective streets key. So we can adress this later.

Nonetheless, we can quickly have a look at the traced roads where the 15K+ measurements where taken, and also visualise the road class.

mapview::mapview(berlin_bike, zcol ="STRKLASSE1")

And now the data-based BBOX:

berlin_bike_bbox <- sf::st_bbox(berlin_bike)

mapview::mapview(list(st_as_sfc(berlin_bike_bbox), berlin_bike))

Easy, right?

2.1.3 Hand-drawn BBOX with `mapedit`

If neither the feature/object-based nor the geodata-based BBOX is sufficient for you (i.e. because you need a custom extent not matching the data), we can literally draw the box with the super-handy mapedit package. For this, we’ll take the cycling data just to add context, and then will interactivly draw a rectangle polygon and store the output in an object.

Don’t forget to click the “Done” button to finish editing.

(berlin_custom <- mapview::mapview(berlin_bike) %>% mapedit::drawFeatures())

# saveRDS(berlin_custom, file = str_c(here::here("data", "GIS_workflows", "/"), "berlin_custom.rds"))
# berlin_custom <- readRDS(file = str_c(here::here("data", "GIS_workflows"), "/", "berlin_custom.rds"))

The resulting object is not a BBOX, obviously, but we can just fetch the object’s BBOX, and voila - habemus BBOX.

berlin_custom_bbox <- sf::st_bbox(berlin_custom)

Let’s double-check our custom BBOX with the cycling data and the Berlin feature object.

mapview::mapview(list(sf::st_as_sfc(berlin_custom_bbox), berlin_sf, berlin_bike))

Dammit, I messed up the lower-right corner by ommiting a piece of Berlin’s south-east (if the use case were to include all of Berlin). But we can easiliy re-draw the box with mapedit or use the bbox <- bbox + vector hack from above.

berlin_custom_bbox

##     xmin     ymin     xmax     ymax 
## 13.02704 52.35716 13.79334 52.67722

We’re after ymin in this case, so we could either do berlin_custom_bbox["ymin"] <- newValue or manipulate the value with berlin_custom_bbox["ymin"] - 0.1.

Ok, enough BBOXing for today. Next: How to get any vector data we want into R.

2.2 Basemap Contents

For creating custom map / visualisation you might want to have as much control of what’s displayed on the basemap level as possible. Depending on the resolution / zoom-level of your map, 1:10m vector data from Natural Earth might already be sufficient. For more detail and features (esp. to play around with when you have a certain Gestalt concept and/or want to map below regional level), querying OpenStreetMap data is probably the gold standard. We’ll look into both approaches in this Chapter.

2.2.1 Quick: Natural Earth

For the regional level, 1:10m vector data might already be sufficient. Let’s look at a Berlin map built from roads, rivers and lakes (and borders, of course). I’m going to load the data from my D:/GIS folder, but you could download everything as described above. Europe is priveledged, so there’s supplemental data offered by Natural Earth for lakes and rivers (cf. Natural Earth: Rivers for more detail and licensing), ~~so I’m going to use it here~~. However, the lakes data does not offer any objects in or around Berlin, so I’m sticking to the regular lakes data instead. For rivers, we get 3 results, but we might want to double-check with the base rivers data. Cause we can ;) And because inspecting your data before working with it is a core skill anyways, GIS or not.

But see for yourself:

Supplemental data for European lakes (on the left) and rivers (on the right) from Natural Earth. Displayed with mapview of course!

So let’s fetch roads, lakes, and rivers from Natural Earth.

roads10 <- rnaturalearth::ne_download(scale = 10, type = "roads", category = "cultural", returnclass = "sf")

# basic lakes data
lakes10 <- rnaturalearth::ne_download(scale = 10, type = "lakes", category = "physical", returnclass = "sf")

# supplemental lakes data for Europe
# lakes10 <- rnaturalearth::ne_download(scale = 10, type = "lakes_europe", category = "physical", returnclass = "sf")
# saveRDS(lakes10, file = str_c(here::here("data", "GIS_workflows", "/"), "lakes10_europe.rds"))
# lakes10 <- readRDS(file = str_c(here::here("data", "GIS_workflows"), "/", "lakes10_europe.rds"))

# basic river data
rivers10_base <- rnaturalearth::ne_download(scale = 10, type = "rivers_lake_centerlines", category = "physical", returnclass = "sf")

# supplemental river data for Europe
rivers10_europe <- rnaturalearth::ne_download(scale = 10, type = "rivers_europe", category = "physical", returnclass = "sf")
# saveRDS(rivers10, file = str_c(here::here("data", "GIS_workflows", "/"), "rivers10_europe.rds"))
# rivers10 <- readRDS(file = str_c(here::here("data", "GIS_workflows"), "/", "rivers10_europe.rds"))

load("D:/GIS/vector/sf/roads10.RData")
load("D:/GIS/vector/sf/lakes10.RData")
load("D:/GIS/vector/sf/rivers10.RData")

Now we’ll only pick those features which are needed aka are contained within our BBOX. There are a couple of approaches (i.e. by subsetting with a toponym, filtering by topological relation between two features (i.e. X within Y, intersects with), or the brute way of cropping with a BBOX).

But before we procedd, let’s inspect the lakes and rivers (base and Europe) objects, as we’re interested in finding out what we got:

mapview::mapview(list(rivers10_base, rivers10_europe, lakes10))

Comparison of base and supplemental rivers data for European. Mapview is just awesome for this kind of exploratory tasks.

See how handy mapview + sf is? We can visually compare two different sets of geometric features with the same easiness / usability and elegance as we do our usual exploratory data analysis of numeric data with summary(), visdat, skimr, ggplot() and so on.

Anyways, we can see from our visual exploration that rivers10_base does not offer any rivers near or in Berlin. We can ignore this object.

Let’s look into subsetting the rest.

2.2.1.1 Subsetting by Toponym / Property

We did this already: We’ve subsetted the berlin_sf object from substates10 with filter(name == "Berlin"). That’s pretty straighforward if your data is clearly subsetable with an adressable object or explicit property (such as type = "highway"). However, that’s usually a bit trickier for cross-boarder features such as roads or rivers.

Deciding on whether to filter (or rather: query) or to crop depends on the size of the set you’re working with. And in practice, a combination of filter and crop is probably the most effective way. Our roads10 object contains 56K features, while lakes and rivers combined are around 3K. Even if 56K is non that much in 2019, you’ll still notice that it takes some time to compute the query (Big O calling!). If you have a dozen or so geometric objects to subset, you’ll quickly realise that size matters and that a clever query/crop/filter approach might save you resources.

Let’s look at sf::st_crop and then at filter(sf::st_intersection()), which probably is not that intuitive when you’re new to GIS/spatial with R.

2.2.1.2 Subsetting by Cropping: `sf::st_crop`

Cropping is straight-forward. We take our BBOX object (or any other feature from which st_crop can derive a BBOX) and then literally cut off all elements that extent beyond the BBOX.

lakes_berlin <- lakes10 %>% sf::st_crop(berlin_custom_bbox)

## although coordinates are longitude/latitude, st_intersection assumes that they are planar

## Warning: attribute variables are assumed to be spatially constant
## throughout all geometries

rivers_berlin <- rivers10_europe %>% sf::st_crop(berlin_custom_bbox)

## although coordinates are longitude/latitude, st_intersection assumes that they are planar

## Warning: attribute variables are assumed to be spatially constant
## throughout all geometries

Now we can simply use mapview or plot the resulting object with ggplot and geom_sf.

mapview::mapview(list(sf::st_as_sfc(berlin_custom_bbox), lakes_berlin, rivers_berlin))

BBOX + cropped rivers and lakes

And now let’s plot the result with ggplot and berlin_sf as the background and use an equal-area projection for Berlin with crs = 3068 with is equivalent to crs = sf::st_crs(3068)

If you need a particular CRS for your use case, https://epsg.io/ might come in handy.

ggplot() +
  geom_sf(data = berlin_sf, fill = "lightgrey", color = "black") +
  geom_sf(data = lakes_berlin, fill = "blue") +
  geom_sf(data = rivers_berlin, aes(size = strokeweig), color = "blue") +
  geom_sf_label(data = rivers_berlin, aes(label = name)) +
  scale_size(range = c(0.5,2)) +
  theme_minimal() +
  guides(size = FALSE) +
  xlab("") +
  ylab("") +
  coord_sf(crs = 3068)

2.2.1.3 Subsetting by Topological Relation

This is as accurate as it gets, but you might need to take some time to understand the basic set of geometric operations. (Actually, I just understood sf::st_intersection while writing this paragraph… .) Just as a handy example, let’s say that we want to reduce our (world) roads10 object to Berlin’s shape. This a regular geocomputational operation, and we’re going to use sf::st_intersection(). This is not the same as sf::st_within(), which will be used next.

As we already know that the result cannot extent beyond Berlin’s BBOX, we should first crop the roads10 object with Berlin’s BBOX in order to reduce the computational cost of this operation. Mind the . placement.

roads_berlin_sf <- roads10 %>% 
  sf::st_crop(berlin_custom_bbox) %>% 
  sf::st_intersection(berlin_sf)

## although coordinates are longitude/latitude, st_intersection assumes that they are planar

## Warning: attribute variables are assumed to be spatially constant
## throughout all geometries

## although coordinates are longitude/latitude, st_intersection assumes that they are planar

## Warning: attribute variables are assumed to be spatially constant
## throughout all geometries

Here’s the result (and Kudos to Ryan Peek “clip object to shape” lingo):

ggplot() +
  geom_sf(data = berlin_sf) +
  geom_sf(data = sf::st_as_sfc(berlin_custom_bbox), alpha = 0.3) +
  geom_sf(data = roads_berlin_sf, aes(color = type)) +
  theme_minimal() +
  coord_sf(expand = FALSE, clip = "on")

I guess it’s pretty obvious how powerful the sf package and geom_sf() are. But let’s not forget about mapview()!

mapview::mapview(list(berlin_sf, roads_berlin_sf))

BBOX + cropped rivers and lakes

Now let’s look at sf::st_within() which calls for a different subsetting logic. The operation returns a binary predicate {0,1}, so we need to filter lengths() > 0. No idea :) Just read `sf``s Vignette. Whatever works, works. In this example, we only keep those roads, which are fully contained within Berlin’s border.

roads_berlin_sf <- roads10 %>% 
  sf::st_crop(berlin_custom_bbox) %>% 
  filter(lengths(st_within(., berlin_sf)) > 0)

## although coordinates are longitude/latitude, st_intersection assumes that they are planar

## Warning: attribute variables are assumed to be spatially constant
## throughout all geometries

## although coordinates are longitude/latitude, st_within assumes that they are planar

ggplot() +
  geom_sf(data = berlin_sf) +
  geom_sf(data = sf::st_as_sfc(berlin_custom_bbox), alpha = 0.3) +
  geom_sf(data = roads_berlin_sf, aes(color = type)) +
  theme_minimal() +
  coord_sf(expand = FALSE, clip = "on")

Of course, at this level of resolution the 1:10m data from Natural Earth does not offer enough granularity (i.e. to represent Berlin solely based on the signatures of Berlin’s roads). Let’s escalate with OpenStreetMaps.

2.2.2 Custom: OSM with `osmdata` (and all the above)

TBD

Multiple non-developer ways to get there:

https://taginfo.openstreetmap.org
https://www.openstreetmap.org
https://overpass-turbo.eu to export as JSON + https://mapshaper.org to turn the JSON layers in question into a condensed shapefile (and simplify the features with a GUI, if necessary)

Here’s a simple query as an example:

berlin_osm_roads <- osmdata::opq(bbox = "berlin") %>% # alternative: bbox = osmdata::bbox_to_string(berlin_bbox)
  # osmdata::add_osm_feature(key = "highway") %>% 
  osmdata::add_osm_feature(key = "highway", value = c("primary", "trunk", "motorway", "secondary", "tertiary")) %>%
  # osmdata::add_osm_feature(key = "highway", value = "primary") %>% 
  osmdata::osmdata_sf()
# saveRDS(berlin_osm_roads, file = str_c(here::here("data", "GIS_workflows", "/"), "berlin_osm_roads.rds"))
# berlin_osm_roads <- readRDS(file = str_c(here::here("data", "GIS_workflows"), "/", "berlin_osm_roads.rds"))

The returning object is a bit messy. It’s a list of lists of simple features (since we used osmdata::osmdata_sf())

berlin_osm_roads %>% names()

## [1] "bbox"              "overpass_call"     "meta"             
## [4] "osm_points"        "osm_lines"         "osm_polygons"     
## [7] "osm_multilines"    "osm_multipolygons"

As we are interested in roads and our query only called for roads as feature, we should focus on osm_lines and osm_multilines.

berlin_osm_roads$osm_lines %>% count(highway) %>% arrange(desc(n))

## Simple feature collection with 15 features and 2 fields
## geometry type:  GEOMETRY
## dimension:      XY
## bbox:           xmin: 13.1191 ymin: 52.34377 xmax: 13.56479 ymax: 52.68006
## epsg (SRID):    4326
## proj4string:    +proj=longlat +datum=WGS84 +no_defs
## # A tibble: 15 x 3
##    highway          n                                              geometry
##    <fct>        <int>                                       <GEOMETRY [Â°]>
##  1 secondary     5918 MULTILINESTRING ((13.2694 52.35693, 13.26932 52.3571~
##  2 tertiary      3171 MULTILINESTRING ((13.41056 52.3576, 13.41102 52.3573~
##  3 primary       2515 MULTILINESTRING ((13.45034 52.36172, 13.45075 52.361~
##  4 motorway       678 MULTILINESTRING ((13.5479 52.36141, 13.54799 52.3608~
##  5 motorway_li~   626 MULTILINESTRING ((13.54776 52.36409, 13.54754 52.363~
##  6 trunk          165 MULTILINESTRING ((13.28164 52.35378, 13.28111 52.354~
##  7 secondary_l~   147 MULTILINESTRING ((13.37566 52.40907, 13.37564 52.409~
##  8 primary_link   125 MULTILINESTRING ((13.36314 52.37029, 13.36333 52.370~
##  9 trunk_link      91 MULTILINESTRING ((13.26929 52.3575, 13.26908 52.3575~
## 10 residential     52 MULTILINESTRING ((13.42551 52.47918, 13.42567 52.479~
## 11 tertiary_li~    47 MULTILINESTRING ((13.51811 52.36478, 13.51822 52.364~
## 12 service         10 MULTILINESTRING ((13.484 52.51135, 13.48426 52.51132~
## 13 unclassified     8 MULTILINESTRING ((13.19101 52.51405, 13.19117 52.514~
## 14 living_stre~     2 MULTILINESTRING ((13.50144 52.51034, 13.50148 52.510~
## 15 path             1 LINESTRING (13.39146 52.50093, 13.3915 52.50065, 13.~

berlin_osm_roads$osm_multilines %>% count(highway) %>% arrange(desc(n))

## Simple feature collection with 2 features and 2 fields
## geometry type:  MULTILINESTRING
## dimension:      XY
## bbox:           xmin: 13.1191 ymin: 52.45264 xmax: 13.50482 ymax: 52.529
## epsg (SRID):    4326
## proj4string:    +proj=longlat +datum=WGS84 +no_defs
## # A tibble: 2 x 3
##   highway       n                                                  geometry
##   <fct>     <int>                                    <MULTILINESTRING [Â°]>
## 1 primary       4 ((13.38425 52.45346, 13.38427 52.45337, 13.38436 52.4528~
## 2 secondary     3 ((13.43335 52.46403, 13.43357 52.46372, 13.43387 52.4633~

berlin_osm_roads_crop <- berlin_osm_roads$osm_lines %>% sf::st_intersection(berlin_sf)

## although coordinates are longitude/latitude, st_intersection assumes that they are planar

## Warning: attribute variables are assumed to be spatially constant
## throughout all geometries

# berlin_osm_roads_crop2 <- berlin_osm_roads$osm_multilines %>% sf::st_intersection(., berlin_sf)

Just to give you an idea: this is the level of granularity (and memory-intense granular vector data) we get:

plot(berlin_osm_roads_crop["highway"])

To be continued…

1 A Bit of a Context